This repository will maintain simulation files and other relevant files of the SAR ADC block worked on in the VSD Summer Online Internship 2020
Note: Circuit requires further optimization to improve performance. Design yet to be modified.
- A Glance at the ADC IP
- Block Diagram of the SAR ADC IP
- Internal Block Diagram of the SAR ADC IP
- Performance Parameters of the SAR ADC
- Subcircuits within the SAR ADC
- Layout
- Running the Simulation
- Pre-Layout Simulation
- Post-Layout Simulation
- Open-Source VLSI Tools
- Issues | Improvements | Future Work
- Acknowledgments
- Contact Information
The Design Specifications of the ADC IP can be found here.
The SAR ADC is built using 5 important blocks:
- Anti-Aliasing Filter
- Sample and Hold
- Comparator
- SAR
- DAC
Parameter | Description | Min | Typ | Max | Unit | Condition |
---|---|---|---|---|---|---|
VDDA | Analog Supply Voltage | 3.2 | V | T=40C to 85C | ||
VDD | Digital Supply Voltage | 1.8 | V | T=40C to 85C | ||
VREFH | Reference Voltage High | 3.3 | V | T=40C to 85C | ||
VREFL | Reference Voltage Low | 0 | V | T=40C to 85C | ||
FCLK | Clock Frequency | 0.01 | 1 | 2 | MHz | T=40C to 85C |
RES | Resolution | 10 | Bits | For all above typical conditions (T=27C) | ||
INL | Integral Non-Linearity | LSB | For all above typical conditions (T=27C) | |||
DNL | Differential Non-Linearity | LSB | For all above typical conditions (T=27C) | |||
RIN | Analog Input Resistance | 110 | kohm | T=-40C - 85C | ||
CL | Analog Input Capacitance | pF | VT=-40C - 85C | |||
IVREF | Current through Reference Source | 1.06 | mA | For all above typical conditions (T=27C) | ||
T1 | Start signal duration | 0.5 | Clock Cycles | T=-40C - 85C | ||
TCONV | Conversion Time | 12 | Clock Cycles | T=-40C - 85C | ||
T4 | Tracking Time | 4 | us | T=-40C - 85C | ||
IDDA | Analog Supply Current | 2.97619 | mA | T=27C, EN=1,FCLK=2MHz | ||
IDDA | Analog Supply Current | pA | T=27C, EN=0,FCLK=2MHz | |||
IDDD | Digital Supply Current | 2.833 | mA | T=27C, EN=1,FCLK=2MHz | ||
IDDD | Digital Supply Current | pA | T=27C, EN=0,FCLK=2MHz |
To enter the Ngspice Shell, open the terminal & type:
$ ngspice
To simulate a netlist, type:
ngspice 1 -> source <filename>.cir
You can exit from the Ngspice Shell by typing:
ngspice 1 -> exit
or
ngspice 1 -> quit
To clone the Repository and download the Netlist files for Simulation, enter the following commands in your terminal.
$ sudo apt install -y git
$ git clone https://github.com/ADC-TEAM2020/SAR_ADC
$ cd SAR_ADC/Simulation/PreLayout_v1.0/
Open the sar_adc.cir file and enter the input voltage [0V-3.3V] as shown in the image below.
Run the netlist file using the following command.
$ ngspice sar_adc.cir
Observe the corresponding Digital Output when EOC[End-of-Conversion Signal] is HIGH between 11.5us -12.0us
Vin = 2000mV
Digital Code= 1001101101 (621)
Max DNL Error = 7.663975155279 LSB
To be recomputed using Best Fit Curve method, accounting for gain and offset error
To be recomputed using Best Fit Curve method, accounting for gain and offset error
Ngspice is an open source mixed-signal circuit simulator.
Open your terminal and type the following to install Ngspice
$ sudo apt-get install -y ngspice
Magic is a VLSI layout tool.
Type the following to install Magic in Ubuntu
$ wget http://opencircuitdesign.com/magic/archive/magic-8.3.54.tgz
$ tar xvfz magic-8.3.54.tgz
$ cd magic-8.3.54
$ ./configure
$ sudo make
$ sudo make install
- Script to automate DNL/INL Calculations completely will be developed.
- Layout will be redone following recommended Mixed Signal Layout practices.
- INL will be recalculated using the Best Fit Curve method, accounting for gain and offset error.
- DNL error for lower digital codes needs to be reduced.
- Sample & Hold must be optimized to improve hold capabilty.
- Comparator design should be modified to reduce Deadband and Offset.
- Opamp to be used in DAC must be optimized to give a better transient response and higher Bandwidth.
- The number of transistors used in a SAR logic must be reduced. The Non-Redundant Successive Approximation Register architecture, which requires only 'N' D flip flops for N-bit conversion, can be employed for this.
- Anti-Aliasing Filter yet to be designed.
- Sheryl Serrao
- Ananya Ghorai
- Shalini Priya
- Uday Vempalli
- Kunal Ghosh
- Kunal Ghosh, Co-founder, VSD Corp. Pvt. Ltd.
- Philipp Gühring, Software Architect, LibreSilicon Assocation
- Saroj Rout, Associate Professor & Chief Mentor of VLSI Center of Excellence SIT, Bhubaneswar, India
- Santunu Sarangi, Asst. Professor, SIT, Bhubaneswar, India
- Tim Edwards, Senior Vice President of Analog and Design at efabless corporation
- Sheryl Serrao, Undergraduate Student, Mumbai University [email protected]
- Ananya Ghorai, Pursuing M.Tech in VLSI Design, IIT(ISM) Dhanbad, [email protected]
- Shalini Priya, M.Tech Embedded System, NIT Jamshedpur [email protected]
- Uday Vempalli, Undergraduate Student,Siddhartha Institute Of Technology,[email protected]
- Kunal Ghosh, Co-founder, VSD Corp. Pvt. Ltd. [email protected]
- Philipp Gühring, Software Architect, LibreSilicon Assocation [email protected]